Computer Science, Fall 2019
MSCS - Masters Student

Shrija Mishra



I am a Graduate Student at Georgia Institute of Techonology, Atlanta specializing in Machine Learning.
As part of Project Tesseare at the Social WellBeing and Dynamics lab, I work on utilizing multimodal data streams with social media to predict and characterize various attributes of physical and mental well-being of individuals. I am interested in working on ethical AI and ML.













PROJECTS





Microsoft - Core Services Engineering

June 2017 - July 2019

Integral role in designing and developing an intelligent data platform which handles ingestion, compute and publish of sales and marketing data across Microsoft. Thus, running a successful business in an effortless manner.

  • Integrated and automated the deployment of the entire ETL pipeline using Azure Databricks, Data Factory and .Net for the ingestion and publish framework that enabled a dynamic and seamless consumer experience in onboarding terabytes of data. Onboarding performance was improved 5x.
  • Implemented a system that aggregates marketing profile and activity data of prospective customers with the help of Azure DataBricks, enabling Microsoft’s marketers target the right consumer at the right time.


Emotion Recognition Through Facial Gestures

August 2016 - May 2017

Proposed and implemented a novel Convolutional Neural Network architecture to identify different emotions and their intensity level in a human face utilizing deep learning techniques.

  • Published in International Conference on Mining Intelligence and Knowledge Exploration ( MIKE 2017).
  • Part of Lecture Notes in Computer Science Book Series ( LNCS, Volume 10682)


Microsoft - Common Data Platform

May 2016 - July 2016

  • Optimized a Commerce BI job to improve its efficiency by 3x and enabling smoother and faster business analytics.
  • Built a system to predict re-enrollment of Microsoft Partners in the target competency for the Microsoft Partners’ Network program using Azure ML and scikit learn, yielding encouraging accuracy of 80%.



Indian Institute of Technology, Kharagpur

May 2015 - July 2015

Implemented a user authentication system with the help of Line Attribute-Based Feature Vectors of fingerprint images using various matching algorithms involving minutiae points and line attributes joining them. This led to a publication on Cryptographic Key Generation using Biometric at ICCII.









© Copyright 2019, Shrija Mishra